Department of Biostatistics, University of Iowa, Iowa City, IA, United States of America
Mini Review
Advancing Dose-Finding Trials: A Review of the Modified Huber Loss Function for Continual Reassessment Methods
Author(s): Fangfang Jiang, Emine O. Bayman and Gideon K. D. Zamba*
Phase I Dose-Finding (DF) trials struggle to balance efficacy and toxicity, often relying on the Traditional Escalation
Rule (TER) or “3+3” design, which is criticized for ignoring dose-toxicity information. Continual Reassessment
Methods (CRM) improve upon TER using Bayesian frameworks but focus mainly on mean posterior toxicity
estimates, neglecting distribution uncertainties. GMHLF combines quadratic and absolute loss functions to handle
deviations from target doses and incorporates unequal penalties for underdoing and overdosing, controlled by a
tolerance level k. Evaluations show GMHLF accurately identifies the Maximum Tolerated Dose (MTD) with fewer
patients than TER and other methods, even when applied to real trial data. The authors suggest exploring
asymmetric GMHLFs and developing user-friendly tools to enhance practical applications.. View more»